Farmers’ conservative behaviour and adapted strategies:
Economic foundation and implications for structural change
Paper submitted for the EAEPE Conference 2009 in Amsterdam
Research Area [M] Social and Cultural Economics
Anne Margarian
Johann Heinrich von Thünen-Institute (vTI)
Institute for farm economics
Bundesallee 50
38116 Braunschweig
Germany
phone: 49-531/596-5140
fax: 49-531/596-5233
mail: [email protected]
Abstract: In real markets, manifold strategies of agents are observed. It is central for further
economic analyses, whether these differences are attributed to exogenous differences or to
endogenously evolving differences in behaviour. With respect to structural change in
agriculture, attempts to explain different paths of development by exogenous economic
conditions resulted in ambiguous results. Therefore, cultural differences have been proposed
as final cause of differentiated developments. In contrast, it is proposed in the present paper
that behavioural covariances evolve endogenously due to strategic interaction on the landmarket; the coincidence of observed attitudes like “conservatism” and strategic behaviour,
expected under certain structural conditions, is demonstrated empirically. Additionally, the
relation between the resulting growth-orientation of farms, exogenous conditions and specific
adoptions of different types of farms are analysed with the help of farm-advisors’ assessments
of farmers’ behaviour in different districts. It is demonstrated that from the imperfect
competition and the resulting endogenously evolving behavioural heterogeneity a high
specifity of structural change evolves. Somehow counter intuitively, it might be the
recognition of this specifity that could prepare for the formulation of a truly general theory of
structural change (in agriculture).
Keywords: social interaction, economic behaviour, path-dependency, agriculture, structural
change
1 Aim of the study
Within the science of economics between its different sub-disciplines a schizophrenic
situation shows up. On the one side, management-science deals with the myriad of strategic
options firms face in real life and industrial economics classifies these strategies into rough
classes of strategies, which subsequently are explained by peculiarities of production and the
market-structure. Neo-classically based macro-economic analyses of the whole economy or
single sectors, on the other side, in order to motivate its mathematical models, require the
existence of unequivocal equilibriums and optimal strategies. These macro-analyses assume
“thick markets” (Makowski and Ostroy, 1995) without room for strategic variations, as they
are explained by management-science and industrial organisation. Such thick markets require
perfect information and therefore markets without price-externalities (ibid.). Makowski and
Ostroy (1995) write:
1
In contrast to this, we will take a “thin-markets” approach. What we mean by this is that we shall try to
avoid the fixed set of roads upon which individuals travel. The aim is to portray a world in which
economic actors are connected not by several main highways, but by a myriad of individual byways of
their own construction” (S. 811).
The authors formally show how, due to the possibility of multiple rational strategies, a
multiplicity of possible alternative markets and therefore the thread of coordination-failures
evolves. In contrast, advocates of the “thick-market-mentality” (ibid.) of general-equilibriumanalyses in order to cope with the empirical fact of existing strategic variances have to claim
that they reflect adoptions towards differing exogenous conditions.
A sector where differing firm-development-strategies have been detected frequently is
agriculture. For Germany, these strategies are described mainly in sociological studies (e.g.,
Patrick et al. (1983), Sinkwitz (2001), Herrmann (1993), Hildenbrand et al. (1992)). Due to
the lack of economic foundation, the respective differences often were explained by assumed
exogenous differences in attitudes, which can be subsumed under the labels “conservative”
vs. “growth-oriented”. Differing dynamics of farm-structure-developments (Huettel and
Margarian, forthcoming) as well as anecdotic evidence 1 and sporadic scientific observations
(e.g., Ohe (1985), Roeder et al. (2006), Goetz and Debertin (2001), Weiss (1999) and Tietje
(2004)) hint on regional differences or covariances in farmers’ strategies. Corresponding to
the identified differences in attitudes, these covariances were sometimes explained by
exogenous regional cultural differences (e.g., Roeder et al. (2006)). Attempts to explain these
behavioural covariances by “hard” exogenous factors, especially regional labour-marketsituations, produced ambiguous results (e.g., Breustedt und Glauben (2007), Glauben et al.
(2006), Weiss (1999), Goetz und Debertin (2001)).
It has been shown, though, that structural development in agriculture might be largely
determined by the initial structural situation due to existing rents of the status-quo and
strategic interaction on the land-market (Huettel and Margarian, forthcoming; Margarian,
2009). Farmers, whose strategies are not strictly growth-oriented (quantity-followers on the
land-market), might be judged as “conservative” by observers while growth-oriented farmers
(quantity-leaders in strategic competition) leave an entrepreneurial impression on observers.
Therefore, these different types of behaviour need not to be attributed to differing exogenous
conditions or cultural differences anymore, but rather represent alternatives that evolve
endogenously from strategic interaction. Beyond these rough classifications in quantityleaders and quantity-followers, depending on specific and sometimes idiosyncratic
circumstances, countless individual strategies of adoption evolve. This, in accordance with
Makowski and Ostroy (1995), is due to the imperfect markets, resulting uncertainty and direct
interaction among farmers.
Under these circumstances, not only determines farmers’ behaviour structural development
but structural development also influences upon farmers’ behaviour. This interdependence
enables self-reinforcing processes and the establishment of different regimes of development.
Accordingly, regional behavioural (“cultural”) differences and the apparent predominance for
1
See for example dlz 2/2007, S. 147; FAZ 126/2007, S. 15; SZ 135/2006, S. V2/4.
2
example of “conservative” farmers might now be open to economic explanations. The impact
of exogenous factors like the regional labour-market-situation on farmers’ strategic choices
then possibly depends on the type of development that has evolved. It is the aim of the present
study to analyse the interdependence of the initial structural situation, farmers’ observed
attitudes, exogenous influences and concrete strategic choices of different types of farms in
the light of the foregone considerations. Therefore, the scattered knowledge of farm advisors
in different German regions has been gathered and made accessible for further analyses.
The paper is structured as follows: Chapter 2 describes the survey among farm-advisors, on
which the analysis is based. In chapter 2.1, from experts’ assessments a single indicator is
constructed that describes farmers’ general attitudes towards farm-development; in chapter
2.1.1 the apparent “conservatism” of famers, which is attributed to strategies of reduced
growth, is explained in a regression-model. The model is constructed based on hypotheses
derived from the theory of strategic competition on the land-market. In chapter 2.2 more
factors, which are extracted from experts’ assessments with a factor-analysis, are presented;
they describe farmers’ strategies, restrictions and opportunities. The relation between general
attitude, non-economic factors, the economic environment and concrete strategic decisions is
analysed in a path-analysis (chapter 2.3); here, technical aspects concerning the applied
structural-equation-model (SEM) are presented briefly in chapter 2.3.1, while chapter 2.3.2
describes the results of the SEM in path-diagrams. Chapter 3 summarises and draws
conclusions.
2 Empirical analysis of farmers’ strategies
Agricultural advisors usually work within a restricted area. Within this area, they are natural
experts on the farmers’ restrictions and opportunities, their strategies and processes of
decision-making. 221 of these advisors from the western counties of Germany attended an
email-survey that consisted of about 120 statements that were to be evaluated on a scale from
1 (corresponding to “in (nearly) no case”) to 7 (“in (nearly) every case”) (figure 1).
Figure 1: Exemplary excerpt from the survey (translated from German)
2 Aims and values of full-time-farmers in my
region
2.1 Farmers here are entrepreneurial oriented: they
invest labour and capital where it is most profitable.
in
(nearly)
no case
1
2
more in every
so
case
less
so
3
4
5
6
7
don't know
0
2.2 Farmers here are rather conservative/ traditionell:
Their main goal is their farm's stability. They avoid
risk and make small steps of growth in existing types
of production.
Source: Own figure
The experts were asked to concentrate on the situation in a district (“Landkreis”, the European
NUTS III level) of high familiarity. Their assessments concerned 145 different districts of
western Germany. Due to the multiple references of different experts to 61 of these districts, a
validation of the assessments was possible. With a variance-analysis, the significance of the
3
variation of the advisors’ assessments between districts as compared to the variance of the
assessments between advisors was determined. Only after dropping assessments for the same
districts that deviated from each other by more than two points, significance of differences in
the assessments between districts was guaranteed for nearly all statements. In this manner,
about 13 % of all statements were judged as deviating values and replaced by missing values.
The same rate of misjudgement is expected for districts covered by only one expert. From the
remaining values, the mean value was calculated for every district in order to reach a univocal
assessment on the districts’ farming sector. The resulting data set consists of one set of
assessed statements for each of the 145 districts.
2.1 An assessment of farmers’ conservative and entrepreneurial attitudes
In the survey, two dimensions classified farmers roughly. One dimension was traditional or
conservative behaviour, which was ascribed towards farmers whose main aim consists in
stabilising their farm, who therefore avoid risk and make small steps of growth. Observers
perceive such a strategy as traditional, but it also corresponds to the strategy of quantityfollowers in strategic competition on the land-market (Huettel and Margarian, forthcoming).
Of the 221 respondents, five did not assess the corresponding statement. In more than half of
the assessments, farmers in the respective districts were classified as rather conservative (five
or more points on the scale), in less than a quarter as rather not conservative (three or less
points on the scale). The remaining assessments were intermediate (four points on the scale).
Even before removing 17 deviating assessments, the differences in the assessment of
conservative behaviour of farmers between the districts according to the variance analysis
were highly significant. Obviously, the dimension of conservative behaviour corresponded to
an existing common mental construct among farm-advisors.
The second dimension for a rough classification of farmers according to their strategic
behaviour was an entrepreneurial attitude, which was ascribed towards farmers, who invest
capital and labour where they are most profitable. Four assessments were missing for the
corresponding statement. The distribution of assessments revealed that obviously this
dimension measured something different from the first dimension. The variance of
assessments in this case was rather low, with 40 % of all assessments concentrated at five
points on the scale. Accordingly, differences between districts in entrepreneurial attitude were
not significant before removing 24 deviating values. While the first dimension determined the
farmers’ attitude with respect to their farm, this second dimension is concerned with general
behaviour. Only if conservatism has a high and entrepreneurial attitude a low assessment, the
farmers in a district are assessed as conservative in every respect. If both received high points
on the scale, the respective farmers are conservative in agriculture but entrepreneurial
otherwise. The case of low values in both dimensions did not occur (figure 2a). Additionally,
if the first dimension received low values while the second was judged as predominant,
farmers’ entrepreneurial behaviour dominates in farming, too. By subtracting the assessments
concerning traditional behaviour from those concerning entrepreneurial attitudes, a single
indicator (“MODERN”) was created that captured all of these nuances. Due to the fact that
the case of low values in both dimensions did not occur, the middle values of MODERN are
non-ambiguous in their interpretation. Therefore, MODERN shows a strong negative
4
correlation with assessments on conservative attitudes (“tradition”) (figure 2b) and a strong
positive correlation with entrepreneurial attitudes (figure 2c). It was therefore successful in
replacing the two seemingly loosely related dimensions (figure 2a) by a single factor.
Figure 2: Scatter-plots of the dimensions of conservatism and entrepreneurial attitude
and between each dimension and the factor MODERN
a)
b)
c)
MODERN
conservative
entrepreneurial
Source: Own figure
Correlation 0.82
No. of
Obs.
137
MODERN
Correlation -0.86
Correlation -0.42
conservative
Ellips
incl.
95%
entrepreneurial
Moreover, this new indicator could remove possible biases in the assessments due to different
interpretations of the scale of different experts. Figure 3 presents the distribution of the new
indicator.
Figure 3: Distribution of “MODERN” across districts for an assessment of
entrepreneurial attitude within and outside farming
General conservative
attitude
Concervative within agriculture,
otherwise entrpreneurial
Entrepreneurial attitude
predominates also in
agriculture
Source: Own figure
Obviously, the judgement of predominance of general conservative behaviour is rather rare;
this underlines the assumption that observed conservative behaviour of farmers is due to
special structural conditions and restrictions of agriculture. The dependence of a conservative
attitude on regional farm-structure and thereby its equivalence with the notion of quantity-
5
adopting behaviour in strategic competition on the land-market (Huettel and Margarian,
forthcoming) will be demonstrated empirically in the following chapter.
2.1.1 Validation of relation between initial farm-structure and growth-orientation
Observers identify conservative behaviour if growth-orientation or maximisation of profits
does not seem to be the overarching goal. Often, a seemingly cooperative behaviour among
farmers might contribute to the assessment, too. In the presence of rents of the status-quo and
alternative options for the utilization of factors, though, the risk-avoiding stabilising
behaviour of reduced growth might be a rational strategy of farm-/firm-development.
Considering game-theoretic approaches, this strategic option can only be stabilised if agents
interact directly in imperfect competition (Margarian, 2009). In agriculture, the kind of direct
competition on the land-market depends on the initial farm-size-distribution and on the
regional dominant production-type as well as on natural conditions. The former determines
the relative competitiveness of those farms that interact, the latter the level of realisable rents
of the status quo as well as realisable rents of growth (Huettel and Margarian, forthcoming).
In the following analysis, whether farmers are judged as conservative shall be explained by
these conditions in order to demonstrate the economic foundation of observed behaviour. The
relevant conditions for those districts included in the analysis are represented by the variables
in table 1.
Table 1: Variables representing the conditions, which determine farmers’ strategic
behaviour (district-level)
Variable
Meaning
GINI79
Inequality in land-distribution
LESSFAVOURED Share of less favoured land
LANDQUALITY
Quality of land
ARABLE
Share of land cultivated by arable farms
MEANSIZE
Mean farm-size
Theoretical
Mean Std.dev.
range
Min
Max
0-1
0.41
0.05
0.33
0.59
0-1
0.50
0.38
0.00
1.00
0 - 100
47.47
10.46
28.62
75.44
0-1
0.30
0.23
0.01
0.91
>0
30.32
12.35
9.46
70.11
COWS
Dairy Cows per 100 Hectares
>= 0
27.95
18.27
0.26
83.56
VALUEADDED
Value added of agriculture per 10 Hectares (1000€)
>= 0
15.79
8.16
4.63
47.62
RENTEDLAND
Share of rented land cultivated by farmers
0-1
0.43
0.10
0.20
0.71
Number of observations: 86
Source: FDZ (1999), Federal Statistical Office (VALUEADDED), and Agricultural Census 1979 (GINI79)
Those 30 % of the districts with the highest values in MODERN, that is, with entrepreneurial
farmers, and those 30 % with the lowest values, that is, with conservative farmers, were
attempted to be separated by the logistic model applied. Therefore, those districts with values
between zero and one 2 in MODERN, characterising farmers that are conservative in
agriculture but entrepreneurial otherwise, were dropped and only 86 observations (districts)
remained in the sample for estimation 3 . The mean-size of farms in land (MEANSIZE)
2
More symmetric cut-off points resulted in rather unequal classes. The score-test (see the following food-note)
showed that in this case the assumption of proportional odds was rejected. Since the scales are arbitrary, the nonsymmetric cut-off-points therefore are justified.
3
A score-test showed that the assumptions for a multinomial model concerning proportional odds were not
justified for MODERN. The proportional-odds assumption could not be rejected, though, for the model with
6
determines the present full-time farms’ necessity for growth (the realisable rents of growth),
while the original farm-size-distribution (GINI79) is expected to have a high impact on the
growth-history of farms in the districts (Margarian, 2009). The growth-history determines the
strategic regime or the growth-orientation of farms in the district. It can be shown that the
dominance of large farms, due to the ambiguity of the Gini-coefficient 4 , reaches its maximum
at a value of about 0.5. Therefore, for the following regression, the indicator has been
transformed with the function “Gini79*(1-Gini79)” (GINIQUADR), which consistently
reaches its maximum (0.25) at a Gini-coefficient of 0.5. The value added of land
(VALUEADDED) contributes positively to possible rents of the status-quo of agriculture. At
the same time, high value-added is often accompanied by high capital- and labour-intensity,
enabling quantity-leadership of single farms who under these circumstances can signal their
will for growth believably (Margarian, 2009). Quality of land (LANDQUALITY) contributes
positively and the share of less favoured land (LESSFAVOURED) negatively to the rents of
growth. The share of rented land (RENTEDLAND) diminishes possible rents of the statusquo. Milk-production (COWS) is characterised by relatively low land-intensity of production.
Additionally, high investments and asset-specifity contributing to higher rents of the statusquo due to sunk costs. Cash-crop farms (ARABLE) on the contrary are characterised by low
capital- and labour-intensity in relation to land. Therefore, in order to grow substantively
accompanied by a growth in labour, a sudden and massive growth in land would be necessary.
Consequently, cash-crop farms often stick to their present form of labour-organisation; this
organisational specifity restricts their growth-potential. This restriction could be interpreted
positively as rents of the status-quo.
The logistic model explains the belonging to those districts, which have been characterised by
their farmers’ conservative behaviour (low values of MODERN). In the estimation,
interactions among variables were controlled since these interactions allow for the test of
conditional hypotheses (Brambor et al., 2006). For a thorough introduction in logistic
regression with interaction-effects, refer for example to Jaccard (2001). In order to make
coefficients in the presence of interactions easily interpretable, variables have been centred on
their mean (Jaccard, 2001, p. 11). Some of the variables have been rescaled to create
comparable values in order to facilitate convergence (table 2). A more complex model with a
significant impact of land-quality has been estimated, too. The more complex model
significantly reduced remaining variance but I refrain from presentation since it does not
change the main rational but complicates interpretation. Considering the test statistics at the
bottom of table 2, the model is highly significant: The differences in experts’ assessments on
farmers’ conservatism in different regions are explainable by the theoretically relevant
variables from secondary statistics. LANDQUALITY is excluded from this restricted model.
The natural conditions are expressed by LESSFAVOURED100, which contributes only
slightly significant to the probability of conservatism in the district.
three classes. The respective multinomial model including the 53 medium-class observations was estimated. The
coefficients closely resemble results of the model presented in table 1.
4
The inequality of land-distribution measured by the Gini-coefficient would be very high in each of either two
cases. The first is observed if an overwhelming number of very small farms were confronted with a neglect able
number of larger farms. The dominance of large farms on the land-market only exists in the second case, though:
if some very large farms on the one side were confronted with mainly small, and few or none intermediate, farms
on the other side.
7
Higher sunk-costs (COWS) or higher transactions costs (ARABLE) and the connected higher
rents of the status-quo reduce growth-orientation and strengthen the assessment of farmers’
behaviour as “conservative”. The opposite holds true if farms are large, probably because the
growth-history determines experts’ assessments in this case.
Table 2: results of the logistic regression explaining prevailing “conservatism” in a
district via economic strategic behaviour
Parameter
Meaning
Intercept
GINIQUADR100
Gini79*(1-Gini79) *100
LESSFAVOURED100 Share of less favoured land*100
ARABLE100
Share of land cultivated by arable farms*100
MEANSIZE
Mean farm-size
COWS
Dairy Cows per 100 Hectares
VALUEADDED
Value added of agriculture per 10 Hectares (1000€)
RENTEDLAND
Share of rented land cultivated by farmers
GINIQUADR100* VALUEADDED
GINIQUADR100* RENTEDLAND
Test statistics:
Likelihood-Ratio test
Score test
Wald test
Pseudo R-square
Coefficients
4478.70 **
188.50 **
0.03 °
0.16 ***
-0.13 **
0.27 ***
1179.60 **
-502.30 **
49.67 **
-21.15 **
<.0001
0.000
0.05
0.62
Probability of error: °: <=10%; *: <=5%; **:<=1%; ***:<=0.1%
Number of observations: 86
Source: FDZ (1999) and Agricultural Census 1979
The remaining three variables and their interactions are significant, too. Where few large
farms initially dominated the land-market (GINIQUADR100) the probability that farms are
assessed as conservative is higher. This observation is in accordance with theoretic
expectations, since in these regions most farmers would behave as quantity-followers on the
land-market and realise reduced growth only. Also in accordance with theoretic expectations
are the impacts of value added per hectare (VALUEADDED) and of the share of rented land
cultivated by farms (RENTEDLAND). While the former has a positive relation to rents of the
status-quo, the latter contributes negatively to the rents. Consequently, farmers are assessed as
conservative with a higher respectively with a lower probability. Finally, the respective
influence of VALUEADDED and RENTEDLAND each is strengthened, where few large
farms dominated the land-market initially (interactions with GINIQUADR100). Where the
land-market is not dominated by single farms, the potential is higher for more presently
competitive farms to establish as quantity-leaders especially under conditions of high assetspecifity. Where the share of rented land is high, farms potential to stabilise their farms on the
land-market is reduced and quantity-leaders can easily oust less competitive farms from
production, resulting in a higher mobility of farmers, accelerated structural change and the
impression of less conservative attitudes.
Accordingly, farmers act conservatively in the observer’s eyes in land-markets with reduced
competition, where the possibility exists to stabilise farms. Here farmers minimise risks by
8
continual but restricted growth in order to secure sustainable realisation of rents of the statusquo. This general “attitude” (or growth-orientation) of farmers therefore can be explained
economically. The concept of cultural differences, a theoretically weak argument anyway in a
competitive environment, is not necessary any more in order to explain the different observed
behaviour of farmers and the resulting different regimes of growth. It remains open, though,
what impact traditions, the economic environment and the interaction of farms play in the
determination of concrete strategic choices of farms. This question is addressed in the
following chapter.
2.2 Assessments of farmers’ opportunities and strategies
While the indicator for the assessment of farmers’ general attitude was constructed by simple
differencing the assessments concerning two separate dimensions from the survey, indicators
characterising specific restrictions, opportunities and strategies of farmers were created by
factor-analysis. Therefore, in a first step those statements of the survey were selected that
were assumed to relate to a common underlying concept. The following concepts are
supposed to characterise the farms’ situation in the districts:
-
strategies of conservative farms,
-
strategies of entrepreneurial farms,
-
general strategic main focus,
-
farmers’ opportunities,
-
reasons for exiting from production and
-
general attitudes.
With the assessed statements of each conceptual group, separately one factor-analysis has
been conducted that usually allowed the identification of about five factors. These factors
describe different dimensions of each concept. The following figure 4 illustrates the logical
flow of the analysis.
Figure 4: Distribution of “MODERN” across districts for an assessment of
entrepreneurial attitude within and outside farming
Conceptual groups...
...of statements
Factors
Strategies of
entrepreneurial
farmers
entreprenneurial farms grow in land
entrepreneurial farms itensify production
entrepreneurial farms cooperate
ENTRGROW
Strategies of
conservative
farmers
conservative farms grow in land
conservative farms work off-farm
conservative farms diversify production
CONSGROW
ENTRINTENS
CONSDIV
Source: Own figure
9
Factor-analysis is based on the correlation between single variables. If correlations are low,
subsuming the respective variables in common factors might not be adequate. The “measure
of adequacy (MSA)” showed, though, that the variables from the experts’ assessments were
suitable. Due to the facts that unbalanced correlation-matrices were used because of many
missing values and that the subjective assessments might be erroneous, it could not be
expected that constructed factors would explain the total variance within the correlationmatrices. Therefore, a main-axis-analysis was conducted rather than an analysis of principal
components (e.g., Backhaus et al., 2003). The necessary priors for expected communalities
were estimated by the technique of “squared multiple correlations (SMC)” (Loehlin, 2004,
p. 161). The creation of factors also served the aim to raise validity of the experts’ subjective
assessments by compensating erroneous assessments of single statements with other
assessments for additional statements. Due to the expected restricted validity of single values
the decision on the number of extracted factors rested not on one formal criterion but rather
by weighing up between Kaiser-criterion, Scree-Test (e.g., Backhaus et al., 2003) and
considerations concerning the factors’ content.. Factors were rotated orthogonal and oblique
(ibid.) in order to reach unambiguously interpretable factors. Factors were interpreted
considering all variables with a loading of 0.3 or higher.
Here only those 14 extracted factors are discussed, which are of relevance for the following
causal analysis. They belong to those six categories of mental constructs, which are necessary
in order to characterise regional opportunities and restrictions as well as farmers’ attitudes and
strategies of growth. The related statements that the experts assessed in the survey are
presented in the appendix (table A1). Three of these factors describe opportunities of farmers
in the district: PROFITAGR describes deficits in regional profitability of agricultural
production. INVESTOPTION describes alternative opportunities to invest in the region within
or outside agriculture. JOB describes the probability of an advantageous non-agricultural
employment, which also reduces willingness to invest in agriculture even though accessibility
to credit might be good.
In addition to MODERN (see above), two further factors were extracted that describe
principal attitudes of farmers: STATUS describes whether being a farmer in the farmers’ eyes
is related to a positive social standing. LIVESTILE describes whether belonging to the
farmers’ community is associated with special values. Three other factors were created, which
summarise possible reasons for farm-exits: JOBEXIT describes whether farmers give up due
to better income-alternatives outside from agriculture and thereby expresses the farmers’ exitmobility; DEFICITEXIT characterises exits that are due to deficits in farms’ efficiency.
STRUCEXIT describes planned farm-exits due to structural reasons. Two factors describe
growth-opportunities of farms in the region: GROWTH describes the general growthorientation of farms in a district and LAND describes the situation on the land-market. The
relations with the statements (table A1 in the appendix) shows, that demand for land is higher,
where the relative competitiveness of milk-production compared to other types of production
is restricted, that is, where land is arable. The last four factors, which are of interest here, deal
with certain specific farm-development-strategies. The first two are farm-developmentstrategies of entrepreneurial farmers: ENTRGROW describes the orientation towards growth
within agriculture of entrepreneurial farmers; ENTRINTENS describes the strategy of growth
10
in agriculture via intensification of production. The last two factors describe strategies of
conservative farmers: CONSGROW describes a concentration of conservative farmers on
growth in land; CONSDIV describes whether farms diversify their income-sources by taking
on non-traditional activities.
The descriptive statistics of factors are summarised in table 3. Remembering that the original
assessments could take on a value between one and seven, most of the factors show an
intermediate mean-value of three to five. “JOB” shows a rather low value, which means that
many experts judge on-farm-activities as advantageous as compared to off-farm-engagement.
Table 3: Factors, their meaning and descriptive statistics
Concept
Opportunities of
farmers in
the district
Principal
attitudes of
farmers
Possible
reasons for
farm-exits
Factor
Meaning
N
Mean
Std.dev.
Min
Max
PROFITAGR
Lack of profitability of
agricultural production
81
4.5
1.4
1.6
8.1
INVESTOPTION
Investment-alternatives within or
outside agriculture
81
3.6
0.9
1.4
5.5
JOB
Probability of an advantagous
non-farm employment
81
1.0
0.9
-1.0
3.4
STATUS
Positive social standing of
farmers
101
3.5
0.9
1.2
5.4
LIVESTILE
Special values of farming
community
101
5.9
0.8
3.6
8.0
JOBEXIT
Exit-mobility of farmers
108
-0.8
0.6
-2.4
0.5
DEFICITEXIT
Exits due to low efficiency of
production
108
5.4
0.9
2.4
7.2
STRUCEXIT
Exits due to structural reasons
108
5.5
0.8
2.9
7.5
Growth-orientation of farms
86
1.3
1.1
-1.5
3.5
Tightness of land-market
81
5.0
1.2
0.3
6.9
Growth-orientation of
entrepreneurial farms
109
3.6
0.9
0.8
6.0
Intensification of production of
entrepreneurial farms
109
3.0
0.9
0.0
4.8
Growth in land of conservative
farms
83
1.5
0.8
-0.6
3.3
83
4.7
0.9
2.6
7.6
GROWTH
Growthopportunities of farms LAND
Strategies of ENTRGROW
entrepreneurial farms ENTRINTENS
Strategies of CONSGROW
conservative farms CONSDIV
Diversification of incomesources of conservative farms
Source: Farm-advisor survey 2007
The prevalence of farm-exits (JOBEXIT) due to advantageous income-alternatives, which is
interpreted here as exit-mobility of farms, is correspondingly low. The general growthorientation and the growth-orientation of conservative farms have low values, too. These
preliminary results support the theoretical assumption that it is less so non-economical
behaviour but more so existing rents of the status-quo that cause reduced mobility.
11
2.3 Path-analysis
Interest now focuses on the relationship between farm-structure and the economic
environment on one side and farmers’ attitudes and exiting-decisions as well as their realised
strategies on the other side. These factors’ relation with each other and with the endogenously
evolving situation on the land-market will be analysed in the following. The analysis relies on
the factors described in chapters 2.1 and 2.2. The construction of the model to be estimated is
guided by the simple logic of causation in structural change depicted in figure 5.
Figure 5: Simple logic of causation in structural change
Exogenous
Endogenous
Economic
environment
Farmers‘
attitudes
Farmers‘
exit-mobility
Growthdynamic
Entrepreneurial
farmers‘
strategies
Conservative
farmers‘
strategies
GINI79
PROFITAGR
INVESTOPTION
JOB
LIVESTILE
STATUS
MODERN
DEFICITEXIT
JOBEXIT
STRUCEXIT
LAND
GROWTH
ENTRINTENS
ENTRGROW
CONSDIV
CONSGROW
Source: Own figure
Conditions exogenous to structural change are placed at the beginning of the causal chain.
Following the chain from the left to the right, the degree of their endogenous character rises
continually. In figure 5 one indicator, GINI79, has been added, which is constructed from
secondary statistics, in order to describe the historical farm-land-distribution among farms.
This indicator is the Gini-coefficient of farm-land-distribution between size-classes of farms
in 1979 (see chapter 2.1.1). A higher initial Gini-coefficient, i.e. higher inequality in farmsizes, causes higher competitiveness of larger farms on the land-market resulting in their
accelerated growth but reduced growth of smaller farms (Margarian, 2009). Because the
pattern of structural change therefore is determined partly by the initial structural situation as
well as by the economic environment, farmers’ long-term-expectation concerning their own
opportunities and the surrounding farmers’ decisions, too, relate to these factors.
Since farmers’ general attitudes result from these long-term-expectations (see chapter 2.1.1),
the attitudes, too, are expected to depend on the exogenous economic environment. Farmers’
attitudes, which relate to each farmer’s expectation concerning the competitors’ future
behaviour and his own opportunities, influence upon their exit-mobility. The exit-mobility in
turn, due to the inevitable relatedness of agricultural production to the non-renewable resource
of land, determines the remaining farms’ growth-dynamic. If realised demand for land is high,
farm-exits might be accelerated further. These opportunities of growth determine the
entrepreneurial farmers’ strategies. Entrepreneurial farmers are possible quantity-leaders
(Huettel and Margarian, forthcoming) or first-movers in decisions. In contrast, conservative
farmers as quantity-followers (ibid.) adopt their strategies after anticipating the
entrepreneurial farmers’ decisions. Therefore, their strategic decision is placed at the end of
the causal chain in figure 3. This causal logic of structural change from long-term conditions
to short-term decisions is assumed for the following estimation of a respective formal model.
It might be contested on the ground of theoretic considerations, but it is principally not open
12
for empirical validation (Pearl, 1998). Against it, the sign and the relative strength of the
estimated causal relations or coefficients are open to empirical testing, once the assumed
causal model is agreed upon.
2.3.1 The structural-equation-model
The relationship between farmers’ fundamental mental disposition, their strategic decisions
and their restrictions and opportunities were then analysed with a path-analysis. The simple
version of a structural-equation-model (SEM) (Bollen, 1989; Loehlin, 2004) applied therein
allows for the differentiated simultaneous formulation of complex hypotheses. By the
estimation of latent variables within these models, factor-loadings might be estimated
simultaneously with covariances and causal coefficients. Due to the complexity of the causal
model in the present application, here the factors have been constructed separately on the cost
of efficiency of estimation. Structural models also allow for feedback-loops as well as for
intervening variables. Direct, indirect and total effects can be calculated. The estimation of bidirected causation comes with the danger of non-identified estimators. Therefore, in the
current application, a recursive model without feedback-loops is estimated.
The basic model is (Bollen, 1989, p.80-81)
y = By + Γx + ζ
(1)
where
B = m*m coefficient matrix
Γ (Gamma) = m*n coefficient matrix
y = m*1 vector of endogenous variables
x = p*1 vector of exogenous variables
ζ (Zeta) = m*1 vector of errors in the equations.
In these structural models with observed variables, it is assumed that the observed variable
resembles exactly the underlying construct without measurement-errors. Due to the recursive
nature of the model in matrix B only coefficients below the diagonal are different from zero.
Further, it is assumed that errors of the simultaneous equations are independent from each
other. Therefore, the matrix of their covariances (Ψ) is diagonal. In contrast, covariances
between all exogenous variables (the economic conditions) are estimated and stored in
matrix Φ. Estimation is supposed to replicate the correlation-matrix, which is the startingpoint of estimation. The correlation-matrix is presented in the appendix (table A2).
In the estimated model, all variables that are placed further left in the causal chain are allowed
to affect all variables that are placed right of them in the causal chain. The resulting model is
saturated with zero remaining degrees of freedom, therefore the replication of the initial
correlations is perfect but an assessment of model-fit is not possible. As stressed before, an
13
assessment of the underlying logic is not the aim of the present study either. The strength and
significance of estimated coefficients might be assessed, though (Pearl, 1998) by estimated
coefficients and standard deviations. All estimated coefficients are presented in the appendix
(table A3). The following interpretation of results concentrates on significant estimators, that
is, coefficients wit a t-value of two or above.
2.3.2 Results of the path-analysis
The results of SEM can be presented in so-called path-diagrams, which facilitate the
communication of assumptions and results of the complex models. Figure 6 presents the
significant causal relations between variables as they are implied by the experts’ assessments.
Figure 6: Causal relations between economic environment, structural situation, farmers’
attitude and strategic decisions
INVESTOPTION
0.4
JOB
-0.35
-0.52 JOBEXIT
0.22
-0.26
-0.24
DEFICITEXIT
-
-0.79 STATUS
0.47 0.36
0.29
ENTRGROW
-0.62 0.38
-0.42
0.39 ENTRINTENS
0.24 0.37
MODERN
0.59
0.24 0.26
-0.32
0.56
LAND
GROWTH
-0.36 -0.41 -0.24
LIVESTILE
0.45
0.27
-0.37
STRUCEXIT
0.27
0.23
-0.34
PROFITAGR
CONSDIV
-0.24
CONSGROW
0.41
GINI79
Source: Own figure
Obviously, an insightful direct interpretation is very difficult due to the model’s complexity.
Therefore, central relations are analysed in an isolated manner in the following paragraphs.
Whether farmers are growth-oriented or not (MODERN) cannot be explained within this
model. Therefore, this general attitude towards farm-development here remains quasiexogenous. MODERN has been explained, though, within a more detailed model relying on
the theoretic foundation of strategic interaction on the land-market (chapter 2.1.1). The lower
the potential and desire for extended demand on the land-market is, due to structural
restrictions, the lower is the farmers’ growth-orientation or entrepreneurial attitude
(MODERN) as perceived by the experts. Further relations between mental constructs are
demonstrated by the path-diagram in figure 7. Here, the top-level arrows show that conditions
of production and the initial farm-structure affect the entrepreneurial attitude (MODERN) as
has been demonstrated in chapter 2.1.1.
14
Figure 7: Self-reinforcing circle of strategic growth-oriented behaviour, perception of
status connected to farming and farming as life-style
Initial farmstructure
Conditions of
production
Entrepreneurial
attitude
-
+
Farming as
lifestyle
-0.79
Loss of status if
exiting farming
0.59
Source: Own figure
A further arrow indicates that the existence of an entrepreneurial attitude is strongly
negatively related with the farmers’ fear from loss of status if exiting farming, as the
respective high negative coefficient shows. On the contrary, the fear from loss of status relates
positively to the perception that farmers perceive farming as a lifestyle, which is characterised
by certain behavioural rules. Since a recursive SEM has been modelled, no empirical support
exists for the dotted arrow. It might be hypothesised, though, that the behavioural rules, which
are connected with farming as lifestyle, pose certain restrictions on entrepreneurial behaviour
in agriculture. The dotted arrow with the second negative relation then closes the circle, which
thereby receives a self-reinforcing character as symbolised by the circular arrow and the plussign in the centre of the circle. These relations give an impression of how economic decisionmakers might adopt their aims to restrictions that determine their strategic behaviour. In this
case, goals and values of farmers would at least be partially endogenous to the economic
process.
Another external factor (see figure 5) that influences farm-development-strategies is the
availability of adequate job-opportunities for potential part-time farmers or exiting farmers
(JOB; see figure 6)). Figure 8 shows the way it influences upon growth-strategies of farms
and compares the influence with that of farmers’ attitudes. The left side of figure 8 shows that
growth-orientation of entrepreneurial farms (ENTRGROW) is affected in various ways.
Good off-farm job opportunities (JOB) reduce farm-exits due to tow efficiency of farms
(DEFICITEXIT). Additional income-sources in the presence of inflexible labour-markets
allow farmers to use residual labour on-farm, which results in very low opportunity-costs of
farm-labour. Farm-exits due to low efficiency of farms support a growth-orientation of
entrepreneurial farms. Its reduction in the presence of good job-opportunities therefore results
in a negative indirect effect of off-farm job-opportunities on the growth-orientation of
entrepreneurial farmers (curved greyish arrow). The positive direct effect of job-opportunities
on growth-orientation offsets by far this indirect effect, though, resulting in a positive neteffect of 0.29 (calculated from 0.36 minus 0.07). Nevertheless, the arrows at the bottom-part
of figure 8 show that a high prevalence of an entrepreneurial attitude among farmers
15
(MODERN) contributes much stronger to the growth-orientation of farmers than the off-farm
job-opportunities. MODERN is strongly determined by the initial and present farm-structure,
though (chapter 2.1.1).
Figure 8: Influence of regional job-opportunities on entrepreneurial and conservative
farmers’ farm-development-strategies
Off-farm jobopportunities for
farmers
-0.24
0.27
Farm-exits
due to
structural
reasons
Farm-exits
due to low
efficiency of
farms
0.29
-0.07
0.36
Growth-orientation of
entrepreneurial farms
0.49
-0.62
Loss of status if
giving up farming
0.06
0.23
Growth-strategy
of conservative
farmers
-0.79
0.07
0.27
Incomediversification of
conservative
farms
-0.37
Entrepreneurial
attitude of farmers
Source: Own figure
Off-farm job-opportunities affect conservative farms’ growth-strategies via their positive
impact upon farm-exits due to structural reasons (STRUCEXIT). These exits relate positive
to income-diversification of conservative farms (CONSDIV) on one side and to growthstrategies of conservative farms (CONSGROW) on the other side. A growing differentiation
among conservative farms in full- and part-time farms should therefore be expected. The
greyish arrows show, though, that the off-farm job-opportunities only exert a low effect on
conservative farms’ growth strategies. This confirms results of other studies on the rather low
impact of regional unemployment-rates on structural change in agriculture. 5 The prevalence
of an entrepreneurial attitude of farmers in the region and thereby the initial and present
farm-structure seems to affect much stronger the conservative farms’ decision for incomediversification.
The next exogenous factor (see figure 5) to be analysed with respect to its impact upon
farmers’ strategies is the farmers’ opportunities to invest in new activities (INVESTOPTION)
in- and outside agriculture (figure 9). According to the experts’ assessments, the perceived
loss of status in case of giving up farming (STATUS) positively relates to investment
opportunities. That might be because in many rural areas farmers are among those inhabitants
5
Glauben et al. (2006) conclude for example from their empirical, regression-based analysis of regional factors
that influence farm-exit-rates „that the declining number of farms is significantly influenced by characteristics of
the farm-sector, whereas characteristics of the non-farm-economy seem to be less important“. In the study of
Goetz and Debertin (2001) the regional unemployment-rate remains insignificant in its influence upon farm-exitrates.
16
who have the capability to realise investments; therefore, in regions with good businessopportunities their status is rather high. Status therefore in a world of imperfect capitalmarkets reflects also the economic advantage in access to capital.
Figure 9: The relation between investment-opportunities and growth-orientation of
farmers on one side and status-consciousness and structural dynamic on the
other
Investment
opportunities in new
activities
0.4
Exit-mobility -0.11
-0.52
0.41
0.22
Loss of status if
giving up farming
-0.79
0.47
-0.14
-0.62
0.5
Growthorientation of
entrepreneurial
farms
Entrepreneurial attitude of
farmers
Source: Own figure
A high relevance of status leads to a reduced exit-mobility (JOBEXIT) and to reduced growthorientation especially of entrepreneurial farms (ENTRGROW) within agriculture. Much
stronger than these negative indirect effects of investment opportunities on exit-mobility and
growth-orientation of entrepreneurial farms via the status-effect, though, is the direct
increasing effect of good investment-opportunities on these factors of structural dynamics.
The indirect strengthening effect of entrepreneurial attitude of farmers (MODERN) on
structural dynamics (exit-mobility and growth-orientation of entrepreneurial farms) is higher
than the net effect of investment opportunities as the greyish arrows in the bottom part of
figure 9 show. Obviously, where the entrepreneurial attitude of farmers is strong, exiting
farming on the other hand has been recognised as a normal entrepreneurial reaction towards
tight land-markets and no or little loss of status is connected with this decision. The resulting
indirect impact upon structural dynamics could possibly explain the ambiguity of results of
empirical studies on exogenous factors’ influences.
Lack of profitability of agriculture (PROFITAGR) is the first exogenous factor (see figure 5)
to be considered that relates directly to agriculture. Its impact on farmers’ strategies is
summarised in figure 10. Low profitability of agriculture causes a reduced growth-orientation
of entrepreneurial farms (ENTRGROW), a low desire for growth in land of conservative
farms (CONSGROW) and consequently a reduced general growth-orientation (GROWTH).
On the other hand, it is also related to a reduced exit-mobility (JOBEXIT). Therefore, a
growth-regime of farm-stabilising strategies is expected. This confirms expectations formed
based on the theory of strategic interaction with respect to the land-market (Margarian, 2009).
Under conditions of low profitability of growth, the low exit-mobility shows that farmers are
reluctant to exit from agricultural production due to various kinds of rents of the status-quo.
17
The low pressure on the land-market if growth-orientation is low allows even more farmers to
stabilise their farms and to realise these rents sustainably. Rents of the status-quo, e.g.,
political rents, therefore might explain why land does not lie fallow as has been predicted for
regions of low agricultural competitiveness repeatedly.
Figure 10: The relation between competitiveness of agriculture in the district and
growth-strategies of farmers
Lack of profitability
of agriculture
-0.26
Exit-mobility
-0.42
-0.34
-0.32
-0.41
Growth-orientation 0.08
of entrepreneurial
General growthfarms
orientation
Growth in land of
conservative farms
0.56
Tight land-market
0.37
-0.08
Entrepreneurial farmers
intensify production
Source: Own figure
The bottom part of figure 10 shows that if general growth-orientation is high on the contrary
and land-markets are tight (LAND) entrepreneurial farmers intensify production
(ENTRINTENS) and partly refrain from a pure strategy of growth in land. This shows that
farmers do adopt their strategies to structural conditions in the course of time resulting in
endogenous developments of regional regimes of growth.
The second sectoral exogenous factor (see figure 5) to be considered is the initial inequality in
land-distribution in 1979 (GINI79). It is the only indicator constructed from secondary datasources, but it does show some significant relations towards experts’ assessments (figure 11).
Where the inequality of land distribution was high initially, natural quantity-leaders on the
land-market existed. These would have grown rapidly, while the smaller farms would have
acted as quantity-followers with reduced growth (Margarian, 2009). Accordingly, as figure 11
shows, the general growth-orientation (GROWTH), that is, the growth-orientation of the
majority of farms is rather low. On the other hand, entrepreneurial farmers intensify their
production (ENTRINTENS) after the periods of rapid growth in land. This direct effect is
only counterbalanced slightly by the indirect effect that a lower growth-orientation of the
majority of farms causes a less tight land market (LAND) and thereby supports growth in land
of competitive farms, i.e., less intensification (greyish arrow).
18
Figure 11: The impact of initial structure on present farm-development-strategies
Inequality in landdistribution in 1979
-0.24
General growthorientation
0.56
Tight landmarket
0.24
-0.05
0.37
Entrepreneurial
farmers intensify
their production
0.41
Conservative
farmers want to
grow in land
-0.24
Incomediversification of
conservative
farmers
Source: Own figure
On the other hand, in the entrepreneurial farmer’s shadow, many conservative farmers had
been able to stabilise their full-time-farm, since none of them could establish as additional
quantity-leader. Consequently, income-diversification of conservative farmers (CONSDIV) is
not a frequent phenomenon in the respective districts. Rather, since they had been forced in a
path of restricted growth in the past, conservative farmers want to grow in land
(CONSGROW) further in order to stabilise their farms. Therefore, the initial farm-structure
and the strategic interaction of farms can explain the disturbing observation, that many farms
state a strong will for growth in land but at the same time, all of them restrict themselves to a
minimum level of growth.
Finally, the impact of the sustainable differences in attitude, which are determined to a high
degree by the initial farm-structure (see chapter 2.1.1) on short-term decisions concerning
farm-development, will be analysed (figure 12). The entrepreneurial attitude of farmers
(MODERN) has a direct negative relation with a perceived loss of status if exiting farming
(STATUS). Via the status-effect, the entrepreneurial attitude of farmers indirectly relates
negatively to the perception of farming as a certain lifestyle (LIFESTYLE). One could
conclude that a lack of entrepreneurial attitude, the fear of losing status and the perception of
farming as a lifestyle measure the same concept. A more thorough analysis of the relations
depicted in figure 12 shows, though, that the status-effect and the perception of farming as a
lifestyle rather allow for a further differentiation of farmers’ attitudes.
The fear from loss of status if exiting farming, which is strongly related to a low
entrepreneurial attitude of farmers, consequently causes a low exit-mobility (JOB) of farmers.
On the other hand, where exit-mobility is low, the general growth-orientation (GROWTH) is
rather high, because growth helps a sustainable stabilisation of farms. The fear from loss of
status is therefore indirectly positive related to stronger general growth-orientation (the left
19
greyish arrow in figure 12). The perception of farming as a lifestyle relates not as strongly to
the entrepreneurial attitude of farmers, and the model explains its value only to a small
degree. In contrast to the fear from loss of status, it has a significant direct negative impact
upon general growth orientation.
Figure 12: The impact of general growth-orientation of farmers and short-term
decisions on farm-development
Entrepreneurial attitude of
farmers
-0.79
Loss of status if
exiting farming
-0.47
0.59
Farming as
lifestyle
-0.52
Exitmobility
-0.32
0.17
-0.36
General
growthorientation
-0.62
0.38
Growth-orientation
of entrepreneurial
farmers
Source: Own figure
The contrary can be observed with respect to the growth-orientation of entrepreneurial
farmers (ENTRGROW). Where the status-effect is high, especially the risk-aversion of
entrepreneurial farmers seems to be higher and their growth-orientation is reduced. That
means that growth-orientation of existing entrepreneurial farms is lower, the lower the
predominance of entrepreneurial farms in a region is. In terms of strategic competition, this
observation can be explained by the remaining ability to grow sufficiently by the many less
competitive farms that act as quantity leaders and the lack of possibility of the few
entrepreneurial farms to oust them out of production (Margarian, 2009).
On the other hand, the higher the fear from loss of status is, the higher is also the perception
of farming as a lifestyle. While lifestyle is negatively related to general growth-orientation, it
relates positively to the growth-orientation of entrepreneurial farmers. These therefore seem
to take advantage from the less competitive behaviour of non-entrepreneurial farmers. Due to
the opposing influence of the fear from loss of status and the perception of farming as a lifestyle on general growth-orientation and the growth-orientation of entrepreneurial farmers the
impact of a lack of entrepreneurial attitude of farmers upon these strategies remains
ambiguous. The question, under what circumstances the perception of farming as a lifestyle
might dominate, cannot be answered, since the relevant factors are not in the model
obviously. Therefore, the possibility remains open that cultural differences have at least some
impact upon farmers’ growth-orientation. The idea of “cultural differences” still has to be
reflected critically, though: Obviously, competitive farmers take their advantage of the
20
reduced competitive behaviour by the majority and are not under the influence of this
“culture”. Therefore, the impression remains that it is mainly the combination of restrictions
and opportunities in the presence of rents of the status quo, which shape regionally
behavioural differences.
3 Results and conclusions
Foregone studies have demonstrated the impact of the initial farm-structure on farmers’
growth-orientation. Consequently, a coincidence of strategic behaviour motivated by the
market with “cultural” differences, identified by external observers, is expected. Cultural
differences in this case are regional differences in farmers’ behaviour. The empirical
validation of this coincidence confirmed the theoretically derived result of endogenously
evolving heterogeneity of farmers’ behaviour in the market. Therefore, concrete strategic
adoptions of farmers not only relate to economic conditions and the present farm-structure but
also to the fundamental growth-strategy that has evolved due to past conditions. The present
paper showed that specific adoptions of farmers towards their environment can be explained
economically within the framework of endogenously evolving heterogeneity.
The experts’ assessment of the fundamental mental disposition of farmers could be explained
with the historically caused distribution of land among farms and the present structure of
production. The assumption that behavioural and cultural differences are exogenous to
economic development is being questioned by these results. At least in part, mental
differences are grounded in historical economic conditions. Fundamental mental differences
respectively their historical causes can be shown to have a lasting influence on farms’
strategic development and the resulting production structure. Moreover, the results show that
in the imperfect land-market, rather subtle mental differences like awareness of status or
traditional values impact upon strategic decisions. While these subtle differences in attitude
closely relate to the economically founded growth-orientation, their primary source remains
open for further analyses. Therefore, conditions influence upon the observed structural
change, which present data on observables alone do not reflect adequately.
The consequences of these results might be rather painful in the eyes of “normal” economists
and intervening policy. Farmers’ reaction to economic circumstances is mediated by initial
conditions and evolving strategic attitudes towards growth and the kind of interaction with
other farms. Therefore, a general optimal strategic choice does not exist from the farmer’s
point of view. Economists in analysing structural change have to take account of specific
regional conditions. Depending on the initial situation, very different dynamic equilibriums
might be realised with different implications for social welfare. A non-ambiguous assessment
of possible welfare-effects of political interventions might not be possible any more.
On the other hand, it has been demonstrated that this specifity of structural change does not
prevent the development of a general theoretic framework of structural development. This
framework, which allows for endogenously evolving behavioural heterogeneity, is
characterised, though, by more degrees of freedom than the special case of neoclassical
theory, which still applies for the case of perfect competition. Somehow counter intuitively, it
21
might finally be the recognition of the specifity of economic developments that paves the way
for a general theory of structural change (in agriculture).
References
Backhaus, K., B. Erichson, W. Plinke und R. Weiber, 2003: Multivariate Analysemethoden.
Eine anwendungsorientierte Einführung. Berlin, Heidelberg: Springer.
Bollen, K. A., 1989: Structural Equations with Latent Variables. New York: John Wiley &
Sons.
Brambor T., W.R. Clark und M. Golder (2006): Understanding Interaction Models:
Improving Empirical Analysis. Political Analysis 14, 63-82.
Breustedt, G. und Th. Glauben, 2007: Driving Forces behind Exiting from Farming in
Western Europe. Journal of Agricultural Economics, Vol. 58, No. 1, 2007, 115-127.
FDZ, 1999: Forschungsdatenzentrum [Centre for scientific use of individual data from official
statistics]. Data of the agricultural census 1999. Kiel.
Gasson, R., 1973: Goals and Values of Farmers. Journal of Agricultural Economics 24 (3), S.
521-537.
Glauben, Th., H. Tietje and Ch. Weiss, 2006: Agriculture on the Move: Exploring Regional
Differences in Farm Exit Rates in Western Germany. Jahrbuch der
Regionalwissenschaft 26: 103-118.
Goddard, E. W., A. Weersink, K. Chen and C.G. Turvey 1993: Economics of Structural
Change in Agriculture. Canadian Journal of Agricultural Economics 41, S. 475-489.
Goetz, S.J. and D.L. Debertin, 2001: Why Farmers Quit: A County-Level Analysis. American
Journal of Agricultural Economics 83, S. 1010-1023.
Herrmann, V., 1993: Handlungsmuster landbewirtschaftender Familien. Bamberg: WVBVerlag.
Hildenbrand, B., K. F. Bohler, W. Jahn, R. Schmitt, 1992: Bauernfamilien im
Modernisierungsprozess. Frankfurt: Campus Verlag.
Huettel, S. and A. Margarian, forthcoming. Structural Change in the West German
Agricultural Sector. Agricultural Economics 40 (Supplementary Issue), 2009.
Jaccard, J., 2001: Interaction Effects in Logistic Regression. Sage University Papers Series on
Quantitative Applications in the Social Sciences, 07-135. Thousand Oaks, CA: Sage.
Loehlin, J.C., 2004: Latent Variable Models. An introduction to factor, path, and structural
equation analysis. London: Lawrende Erlbaum Associates.
Makowski, L. und J. M. Ostroy, 1995: Appropriation and Efficiency: A Revision of the First
Theorem of Welfare Economics. The American Economic Review 85 (4), S. 808-827.
22
Margarian, A., 2009: The regional specifity of structural change in agriculture: An assessment
of the role of farmers’ strategic behaviour on the land market. DFG-ResearchGroup 986: SiAg-Working Paper.
http://www.agrar.hu-berlin.de/struktur/institute/wisola/fowisola/siag/publikationen
Ohe, H. von der, 1985: Landwirte im Entscheidungsprozeß – Verhaltenswissenschaftliche
Ansätze und Ergebnisse einer empirischen Untersuchung. Frankfurt: Verlag Alfred
Strothe.
Patrick, G. F., B. F. Blake and S. H. Whitaker, 1983. Farmers’ goals: Uni- or MultiDimensional? American Journal of Agricultural Economics 65 (2), S. 315-320.
Pearl, J., 1998: Graphs, Causality, and Structural Equation Models. Sociological Methods and
Research 27(2), 226-284.
Roeder, N., J. Kantelhardt and M. Kapfer, 2006: Impact of the CAP reform on small-scale
grassland regions. Agrarwirtschaft 55, Heft 5/6, S. 257-267.
Sinkwitz, P., 2001: Landwirtschaftliche Nebenerwerbsbetriebe im wirtschaftlichen und
sozialen Wandel. Land-Berichte 7, 6 (2), Special Issue. Ulrich Planck zum 80.
Geburtstag. Hrsg.: Franz Kromka und Gerd Vonderach.
Tietje, H., 2004: Hofnachfolge in Schleswig-Holstein. Dissertation zur Erlangung des
Doktorgrades der Agrar- und Ernährungswissenschaftlichen Fakultät der ChristianAlbrechts-Universität zu Kiel.
Varian, H.R., 1992: Microeconomic Analysis. New York: W. W. Norton & Company.
Weiss, Ch. R., 1999: Farm Growth and Survival: Econometric Evidence for Individual Farms
in Upper Austria. American Journal for Agricultural Economics 81, S. 103-116.
23
Appendix
A1)
Factors and related statements that were assessed by the experts
Factor
PROFITAGR
INVESTOPTION
JOB
STATUS
LIVESTILE
JOBEXIT
DEFICITEXIT
STRUCEXIT
GROWTH
LAND
ENTRGROW
ENTRINTENS
CONSGROW
CONSDIV
Associated statements and their relation with the factor
+
“Without investment-support profitability of arable farming would be questionable.”
+
“Without investment-support profitability of pig- and poultry-production would be questionable.”
+
“Without investment-support profitability of horticulture would be questionable.”
+
“Without investment-support profitability of milk-production would be questionable.”
+
“Farmers should invest more in new types of production.”
+
“Full-time-farms have additional sources of income apart from their farm”.
“If farmers invest too little it is due to risk-aversion of banks.”
+
“If a farmer here looks for non-farm-employment he will be successful.”
+
“Exiting from farming implies for farmers a loss in status.”
+
“If necessary, farmers support each other based on trust."
“Farmers’ families are willing to sacrifice short-term profits or consumption, in order to secure
+
future stability of the farm."
+
“Income from off-farm-labour is also spent for the financial stabilisation of the farm."
+
“If off-farm-earnings per hour exceed income from farming, farmers exit the sector."
+
“Exiting farms are characterised by low labour-efficiency."
+
“Exiting farms are characterised by low management-ability."
+
“Exiting farms are characterised by deficits in investments."
+
“Exiting farms occupy less land than surrounding farms."
+
“Famers of exiting farms are characterised by good opportunities for non-farm-earnings."
+
“Exiting farms are characterised by deficits in investments."
+
“Farms exit in the course of farm-succession."
+
“Demand for land is higher than supply of land."
+
“In large arable farms in the district non-family-labour is occupied."
+
“Farms in the district cooperate contractually with each other."
+
“Demand for land is higher than supply of land."
“The necessary employment of non-family-labour represents a restriction on growth, because the
+
necessary rapid accompanying growth is not possible."
“Change to part-time-farming is a reaction towards missing possibilities for growth due to the tight
+
land-market."
“Profitability of necessary investments in milk-production would be questionable without
+
investment-aid."
+
“Non-exiting entrepreneurial full-time farms follow the strategy of growth in land."
“Non-exiting entrepreneurial full-time farms follow the strategy of offering services, which are
+
related to agriculture."
+
“Non-exiting entrepreneurial full-time farms follow the strategy of cooperation with other farms."
+
“Non-exiting entrepreneurial full-time farms follow the strategy of growth unrelated to land."
+
“Non-exiting entrepreneurial full-time farms follow the strategy of specialisation.”
“Non-exiting entrepreneurial full-time farms follow the strategy of offering services, which are
related to agriculture."
+
“Non-exiting conservative full-time farms follow the strategy of growth in land.”
“Non-exiting conservative full-time farms follow the strategy of growth unrelated to land."
“Non-exiting conservative full-time farms follow agricultural niche-strategies such as ecological
+
production, breeding etc.”,
“Non-exiting conservative full-time farms follow the strategy of non-agricultural diversification of
+
their farm."
“Non-exiting conservative full-time farms follow the strategy of diversifying income via off-farm+
work, for example by spouse or farm-successor."
“Non-exiting conservative full-time farms follow the strategy of changing their organisation
+
towards part-time-farming."
“Non-exiting conservative full-time farms follow the strategy of offering services, which are
+
related to agriculture."
Source: Own table
24
A2)
Correlation-matrix of variables from the structural-equation-model
5
1
2
3
4
6
7
8
1.00
CONSDIV
1
83
-0.13
1.00
2 CONSGROW
83
83
-0.19
0.14
1.00
3 ENTRINTENS
74
74
109
0.11
0.26
-0.09
1.00
4 ENTRGROW
74
74
109
109
0.16
0.26
0.00
0.15
1.00
5 STRUCEXIT
75
75
92
92
108
-0.11
0.05
0.29
0.01
-0.05
1.00
LAND
6
54
54
73
73
70
81
-0.21
-0.09
0.11
-0.05
-0.09
0.51
1.00
GROWTH
7
56
56
81
81
72
68
86
-0.04
0.00
0.06
0.09
-0.12
-0.09
-0.04
1.00
8 DEFICITEXIT
75
75
92
92
108
70
72
108
0.19
0.03
-0.11
0.04
-0.04
-0.13
-0.11
-0.36
JOBEXIT
9
75
75
92
92
108
70
72
108
0.04
0.17
-0.02
0.26
0.43
-0.03
-0.20
-0.05
10 LIVESTILE
67
67
90
90
88
69
74
88
-0.02
0.02
-0.02
0.03
0.25
0.08
-0.05
0.12
STATUS
11
67
67
90
90
88
69
74
88
-0.09
0.03
0.12
-0.06
-0.19
-0.09
0.05
-0.02
MODERN
12
82
82
107
107
108
80
86
108
-0.22
0.36
0.19
-0.02
0.03
-0.12
-0.21
-0.07
GINI79
13
83
83
108
108
108
81
85
108
0.14
-0.10
0.00
-0.15
0.15
-0.04
-0.36
0.05
14 PROFITAGR
54
54
73
73
70
81
68
70
0.15
0.11
-0.17
0.22
0.05
0.02
-0.08
0.01
15 INVESTOPTION
54
54
73
73
70
81
68
70
-0.11
0.12
0.13
0.21
0.33
0.12
0.18
-0.27
JOB
16
54
54
73
73
70
81
68
70
Remark: top-line: Pearson’s correlation-coefficients; bottom-line: number of observations
Source: Own table
9
10
11
12
13
14
15
16
1.00
108
-0.20
88
-0.28
88
0.20
108
-0.05
108
-0.16
70
0.14
70
-0.01
70
1.00
101
0.66
101
-0.55
101
-0.12
100
0.09
69
0.29
69
0.11
69
1.00
101
-0.82
101
-0.07
100
0.00
69
0.30
69
0.06
69
1.00
137
0.12
136
0.07
80
-0.11
80
-0.14
80
1.00
144
0.18
81
0.05
81
0.07
81
1.00
81
0.31
81
-0.02
81
1.00
81
-0.21
81
1.00
81
25
A3)
Estimated causal coefficients from the structural-equation-model
influences
upon...
CONSDIV
CONSGROW ENTRINTENS ENTRGROW STRUCEXIT
LAND
GROWTH
DEFICITEXIT
JOBEXIT
LIVESTILE
STATUS
MODERN
Coeff t-value Coeff t-value Coeff t-value Coeff t-value Coeff t-value Coeff t-value Coeff t-value Coeff t-value Coeff t-value Coeff t-value Coeff t-value Coeff t-value
CONSDIV
CONSGROW
-0.11
-0.9
ENTRINTENS
0.01
0.1
0.03
0.3
ENTRGROW
0.13
1.1
0.05
0.5
-0.13
-1.1
STRUCEXIT
0.27 *
2.3
0.23 *
2.2
-0.06
-0.5
0.04
3.2
0.13
LAND
-0.02
-0.2
0.21
1.8
GROWTH
-0.17
-1.3 -0.17
-1.4
DEFICITEXIT
-0.03
-0.3
0.15
1.3
0.17
1.4
0.10
0.8
0.02
0.2
0.05
0.4 -0.06
-0.17
-1.1
0.24
1.6
0.22
1.4
0.21
0.9 -0.62 *
-3.2
0.10
0.5
0.08
0.4
0.09
0.4
0.31
1.6 -0.52 * -2.6
0.39 *
2.2 -0.19
-1.1
0.12
0.7 -0.07
-0.4
0.10
0.6
0.16
0.9 -0.18
-1.0 -0.03
-0.2 -0.79 *
0.01
-0.3 -0.24 *
-1.0 -0.05
-0.5 -0.10
-1.2
JOBEXIT
LIVESTILE
STATUS
-0.29
-1.3 -0.23
-1.1
MODERN
-0.37 *
-2.0
0.04
0.2
0.41 *
4.1
GINI79
-0.24 *
-2.0
PROFITAGR
0.10
0.8 -0.34 * -2.8
INVESTOPTION
0.15
1.1
0.18
-0.15
-1.2
0.07
JOB
0.37 *
0.4
-0.05
0.24 *
-0.4 -0.23
0.29 *
0.38 *
1.2 -0.06
-0.6
0.05
0.4
-1.8
0.56 *
4.9
2.8 -0.01
-0.1 -0.12
-1.1 -0.11
-1.0
0.09
0.8 -0.06
-0.5 -0.32 *
-2.9
-0.35 * -3.4
0.45 *
3.2 -0.05
-0.4 -0.36 *
-2.8
-0.23
-0.5
2.7
-1.7 -0.07
-0.5
0.59 *
3.9
-12.9
0.1
0.04
0.4 -0.04
-2.5
-0.10
0.01
0.2
0.12
1.1
-0.04
-0.3 -0.42 * -3.8
0.16
1.4
0.17
1.5 -0.41 * -4.0
0.02
0.2 -0.26 * -2.4
0.08
1.0 -0.02
-0.3
0.10
0.9
1.4
-0.21
-1.5
0.47 *
4.1 -0.10
-0.8
0.00
0.0
0.24 *
2.1
0.00
0.0
0.40 *
3.4
0.11
1.2
0.22 *
3.4 -0.19
-1.6
0.6
0.15
1.2
0.36 *
3.4
2.6 -0.02
-0.2
0.26 *
2.6
0.09
0.8
0.10
1.2
0.00
0.0 -0.18
-1.6
2.3
0.27 *
-0.24 * -2.4
Source: Own table
26